Expert's Angle: Analytics and the Future of Contact Centers: Voice Analytics

What changes are in store for the contact center in terms of analytics? This article series explores some of the major trends that will impact contact centers of all types over the next ten years. Read the previous article.

Voice analytics is one of the most exciting and rapidly developing areas in the contact center right now, and it will continue to see significant growth over the next several years. This area of business intelligence can provide real-time information to agents, as well as important factors to be considered for follow up. It also has the potential to be used in several different call center settings and may even be used to develop new queuing approaches, such as call stress and intent-based cues. The ability to determine the stress of the caller can be used to determine how to route a call. The ability to determine intent based on verbal cues and tone can also be used to help predict several different types of propensities, the propensity to buy in a sales environment and the propensity to pay in a collections environment are but a few. We will discuss each of these approaches briefly in this article as the next topic of how analytics will shape future approaches to contact centers.

Intent-Based Analytics

For the intent-based voice analytics, we will use the example of a collections center, however the same principles for any sales or a medical treatment center can also be implied. Basically, any contact center that conducts outbound or inbound activity that requires the caller or respondent to perform an action based on the outcome of the call can be measured using Voice Analytics. While Voice Analytics is not 100% accurate, it is an excellent lead indicator as to whether or not the caller/respondent will follow through with the action required at the end of the call.

For example, in financial institutions, one of the key areas of interest for Voice Analytics is the intent of the caller. Ever since collections evolved as an activity conducted over the phone, the ability to determine with any sort of reasonability the true intent of the respondent has been limited. Resultant “promises to pay” from caller/respondent campaigns require the financial institution to wait until after the date and time of the promise has expired to determine if the call/respondent has actually followed through with a payment. By this time, expensive and time-consuming methods, such as outbound call attempts, must be performed so as to re-initiate the payment process or reinforce the importance of payment.

With the advent of Voice Analytics, financial institutions can evaluate the tone and intent of the caller/respondent and apply call treatments according to actual propensity-to-pay models developed from this type of analytics. These models allow the contact center to use Voice Analytics as a means to increase compliance and decrease cost based on the appropriate contact channel. For example, financial institutions will have the choices to:

1. Do nothing, as the tone sounded sincere and there was a high likelihood to pay
2. Send a text message or virtual agent message to the caller as a reminder on the day of the promised payment to increase payment compliance
3. Create a full outbound campaign with live agents to perform the reminders for those callers/respondents that had a low propensity or likelihood of payment or compliance

These examples are only one application for Voice Analytics. This same concept can be used in a sales environment, in a medical treatment environment for distance treatment practices or in a loyalty environment to predict churn. Another example of how voice analytics can be used is in how a contact center routes their calls.

IVR Response Queuing

One of the cutting-edge approaches to using Voice Analytics is to take the stress level of the caller - as interpreted through the IVR application - and then route the call based on that stress level. Those callers that show higher stress levels in their voice can either be routed earlier in the queue or routed to a senior agent pool so it can be handled the call more appropriately, and potentially prevent it from becoming a complaint call. This proactive approach is a significant change from the typical visible and invisible queues that are part of the contact center world.

Voice Analytics can also be used to determine queuing by providing an interval-based response determination for an inbound queue. IVR responses can be laced into these interval prompts to gauge the frustration levels of the caller, similar to the spaced responses and prompts callers hear while they await response from an agent in queue. These responses also determine if the call can either be moved in the queue or be routed to the senior agent pool.

Quality Model Voice Analytics

The quality model is another aspect of Voice Analytics that can be used to evaluate agent performance. By analyzing the agent’s performance through a traditional quality model as well as their intent through Voice Analytics, the business will be able to better evaluate the true impact and strategy adherence of the contact center. By measuring the agent’s vocal inflections, as well as the overall tone of the call, the contact will be better equipped to coach agents on how to increase their effectiveness as well as how to manage call control more efficiently.

Modern Voice Analytics and its uses are only limited to how individual contact centers choose to implement the new technology. While earlier versions of this tool did not necessarily deliver on their promise during the earlier part of the past decade, the new advances that measure more of the tone and the intent of the caller promises to deliver new capabilities for the call center of the future. As we begin to find ways and means to customize our contact channel management to the individual caller, these types of tools will become more important in the future.

TIn our next article we will discuss multi-modal contact strategies that further the idea of a customized and unique caller experience based on the analytical approaches that allow for a better understanding of caller/respondent’s preferences and needs.